Having loaded the receiver with the weight I, something heavier than may be necessary to force it through the mercury, it is balanced by the small weight K, which hangs from that part of the spiral where the radius is equal to that of the wheel G, from this point the radius of the spiral must be increased in such proportion, that in every part of its circuit, the weight K may be an exact counterpoise to the airholder. In this way, so little friction will be produced, that merely plunging the lower orifice of the tube D under mercury contained in the small vessel L, will be sufficient to overcome every resistance, and to force the gas discharged from the beak of a retort into the receiver, where whatever may be its quantity, it will be subjected to a pressure exactly corresponding to that of the atmosphere. The edge of the wheel H being graduated, the balance cord _c_ may be made to indicate its volume. Should it at any time be necessary to reduce the pressure to the medium standard of the barometer, it may easily be done by graduating the lower end of the tube D, and adding to the weights I or K, as may be found necessary; the surface of the mercury in the tube pointing out the increase or diminution. The concavity at the top of the internal cylinder is intended to contain any liquid it may be thought proper to expose to the action of the gas. The upper orifice _f_, with its ground stopper, is particularly useful in conveying air from the retort _g_, with its curved neck, into the receiver, without its passing through the tube D. In all cases where a rapid extrication of gas is expected the retort _g_, should be firmly luted to the orifice and the weight I, removed from the top of the receiver, this by diminishing the pressure, will admit the gas to expand freely in the airholder at the instant of its formation, and prevent an explosion of the vessels. The same caution must be observed whenever any inflammation of gas is produced by the electric spark. The air may be readily transferred through water or even mercury by the tube _h_, Fig. 1. To prevent an absorption of mercury in case of a condensation taking place in the retort made use of for generating air, Mr. Davy has applied the stop-cock _i_, to which the neck is firmly luted. This stop-cock is likewise of great service in saturating water with acid or alkaline gases, which may be effected by luting one end of the tube _k_ to the stop-cock, and plunging the other into the fluid in the small vessel _l_, cemented at top, and terminating in the bent funnel _m_—the tube _h_ having been previously removed, and the lower orifice of the tube D either sunk to a considerable depth in mercury, or closed with a ground stopper. The bend of the funnel _m_, may be accurately closed by the introduction of a few lines of mercury. The application of the stop-cock _n_, has enabled Mr. Davy to perform some experiments on respiration with considerable accuracy.